roboflow-100-benchmark
yolov5
roboflow-100-benchmark | yolov5 | |
---|---|---|
8 | 129 | |
227 | 47,071 | |
4.0% | 1.5% | |
0.6 | 8.8 | |
6 months ago | 4 days ago | |
Jupyter Notebook | Python | |
MIT License | GNU Affero General Public License v3.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
roboflow-100-benchmark
-
AI That Teaches Other AI
> Their SKILL tool involves a set of algorithms that make the process go much faster, they said, because the agents learn at the same time in parallel. Their research showed if 102 agents each learn one task and then share, the amount of time needed is reduced by a factor of 101.5 after accounting for the necessary communications and knowledge consolidation among agents.
This is a really interesting idea. It's like the reverse of knowledge distillation (which I've been thinking about a lot[1]) where you have one giant model that knows a lot about a lot & you use that model to train smaller, faster models that know a lot about a little.
Instead, you if you could train a lot of models that know a lot about a little (which is a lot less computationally intensive because the problem space is so confined) and combine them into a generalized model, that'd be hugely beneficial.
Unfortunately, after a bit of digging into the paper & Github repo[2], this doesn't seem to be what's happening at all.
> The code will learn 102 small and separte heads(either a linear head or a linear head with a task bias) for each tasks respectively in order. This step can be parallized on multiple GPUS with one task per GPU. The heads will be saved in the weight folder. After that, the code will learn a task mapper(Either using GMMC or Mahalanobis) to distinguish image task-wisely. Then, all images will be evaluated in the same time without a task label.
So the knowledge isn't being combined (and the agents aren't learning from each other) into a generalized model. They're just training a bunch of independent models for specific tasks & adding a model-selection step that maps an image to the most relevant "expert". My guess is you could do the same thing using CLIP vectors as the routing method to supervised models trained on specific datasets (we found that datasets largely live in distinct regions of CLIP-space[3]).
[1] https://github.com/autodistill/autodistill
[2] https://github.com/gyhandy/Shared-Knowledge-Lifelong-Learnin...
[3] https://www.rf100.org
- Roboflow 100: A New Object Detection Benchmark
-
[R] Roboflow 100: An open source object detection benchmark of 224,714 labeled images in novel domains to compare model performance
I'm Jacob, one of the authors of Roboflow 100, A Rich Multi-Domain Object Detection Benchmark, and I am excited to share our work with the community. In object detection, researchers are benchmarking their models on primarily COCO, and in many ways, it seems like a lot of these models are getting close to a saturation point. In practice, everyone is taking these models and finetuning them on their own custom dataset domains, which may vary from tagging swimming pools from Google Maps, to identifying defects in cell phones on an industrial line. We did some work to collect a representative benchmark of these custom domain problems by selecting from over 100,000 public projects on Roboflow Universe into 100 semantically diverse object detection datasets. Our benchmark comprises of 224,714 images, 11,170 labeling hourse, and 829 classes from the community for benchmarking on novel tasks. We also tried out the benchmark on a few popular models - comparing YOLOv5, YOLOv7, and the zero shot capabilities of GLIP. Use the benchmark here: https://github.com/roboflow-ai/roboflow-100-benchmark Paper link here: https://arxiv.org/pdf/2211.13523.pdf Or simply learn more here: https://www.rf100.org/ An immense thanks to the community, like this one, for making it possible to make this benchmark - we hope it moves the field forward! I'm around for any questions!
-
Introducing RF100: An open source object detection benchmark of 224,714 labeled images across 100 novel domains to compare model performance
Or simply learn more: https://www.rf100.org/
-
We took YOLOv5 and YOLOv7, trained them on 100 datasets, and compared their accuracy! 🔥 The results may surprise you.
github repository: https://github.com/roboflow-ai/roboflow-100-benchmark blogpost: https://blog.roboflow.com/roboflow-100/ arXiv paper: https://arxiv.org/abs/2211.13523
-
Show HN: Real-World Datasets for Benchmarking Object Detection Models
Github: https://github.com/roboflow-ai/roboflow-100-benchmark
At Roboflow, we've seen users fine-tune hundreds of thousands of computer vision models on custom datasets.
We observed that there's a huge disconnect between the types of tasks people are actually trying to perform in the wild and the types of datasets researchers are benchmarking their models on.
Datasets like MS COCO (with hundreds of thousands of images of common objects) are often used in research to compare models' performance, but then those models are used to find galaxies, look at microscope images, or detect manufacturing defects in the wild (often trained on small datasets containing only a few hundred examples). This leads to big discrepancies in models' stated and real-world performance.
We set out to tackle this problem by creating a new set of datasets that mirror many of the same types of challenges that models will face in the real world. We compiled 100 datasets from our community spanning a wide range of domains, subjects, and sizes.
We've benchmarked a couple of models (YOLOv5, YOLOv7, and GLIP) to start, but could use your help measuring the performance of others on this benchmark (check the GitHub for starter scripts showing how to pull the dataset, fine-tune models, and evaluate). We're very interested to learn which models do best in which real-world scenarios & to give researchers a new tool to make their models more useful for solving real-world problems.
yolov5
-
จำแนกสายพันธ์ุหมากับแมวง่ายๆด้วยYoLoV5
Ref https://www.youtube.com/watch?v=0GwnxFNfZhM https://github.com/ultralytics/yolov5 https://dev.to/gfstealer666/kaaraich-yolo-alkrithuemainkaartrwcchcchabwatthu-object-detection-3lef https://www.kaggle.com/datasets/devdgohil/the-oxfordiiit-pet-dataset/data
- How would i go about having YOLO v5 return me a list from left to right of all detected objects in an image?
-
Building a Drowsiness Detection Web App from scratch - pt2
!git clone https://github.com/ultralytics/yolov5.git ## Navigate to the model %cd yolov5/ ## Install requirements !pip install -r requirements.txt ## Download the YOLOv5 model !wget https://github.com/ultralytics/yolov5/releases/download/v6.0/yolov5s.pt
-
[Help: Project] Transfer Learning on YOLOv8
Specifically what I did was take the coco128.yaml, added 6 new classes from Dataset A (which have already been converted to YOLO Darknet TXT), from index 0-5 and subsequently adjusted the indices of the other COCO classes. The I proceeded to train and validate on Dataset A for 20 epochs.
-
Changing labels of default YOLOv5 model
I am using the default YOLOv5m6 model here with sahi/yolov5 library for my object detection project. I want to change just some of labels - for example when YOLO detects a human, I want it to label the human as "threat", not "person". Is there any way I can do it just changing some code, or I should train the model from scratch by just changing labels?
-
First time working with computer vision, need help figuring out a problem in my model
You should add them without annotations. Go through this.
-
AI Camera?
You are correct and if you check the firmware, it's yet another famous 3rd party project without attribution, namely https://github.com/ultralytics/yolov5
-
First non-default print on K1 - success
On one side, being a Linux user for 24 years now, it annoys me that they rip off code and claiming it as theirs again, thus violating licenses, but on the other thanks to k3d's exploit I'm able to tinker more with the machine and if needed do (selective) updates by hand then with a closed source system. It's not just "klipper", with klipper, fluidd and moonraker, it's also ffmpeg and mjpegstreamer. It's gonna be interesting since they also use a project that isn't just GPL, but APGL (in short "If your software gives service online, you have to publish the source code of it and any library that it borrows functions from.") - they use yolov5 (for AI).
- How does the background class work in object detection?
What are some alternatives?
Shared-Knowledge-Lifelong-Learnin
mmdetection - OpenMMLab Detection Toolbox and Benchmark
Shared-Knowledge-Lifelong-Learning - [TMLR] Lightweight Learner for Shared Knowledge Lifelong Learning
detectron2 - Detectron2 is a platform for object detection, segmentation and other visual recognition tasks.
make-sense - Free to use online tool for labelling photos. https://makesense.ai
darknet - YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )
roboflow-100-benchmark - Code for replicating Roboflow 100 benchmark results and programmatically downloading benchmark datasets [Moved to: https://github.com/roboflow/roboflow-100-benchmark]
Deep-SORT-YOLOv4 - People detection and optional tracking with Tensorflow backend.
fasterrcnn-pytorch-training-pipeline - PyTorch Faster R-CNN Object Detection on Custom Dataset
yolor - implementation of paper - You Only Learn One Representation: Unified Network for Multiple Tasks (https://arxiv.org/abs/2105.04206)
autodistill - Images to inference with no labeling (use foundation models to train supervised models).
OpenCV - Open Source Computer Vision Library